The Classification of Ordinal Data for Choropleth Mapping
Robert G. Cromley, Richard D. Mrozinski · The Cartographic Journal · 1999
Choropleth maps have traditionally displayed standardized interval or metric data for areal units. The data classifications for such maps are usually based on an attribute's statistical distribution without regards for its spatial distribution. This paper presents a classification scheme that focuses on an attribute's spatial arrangement without regard to its statistical distribution. Using only its ordinal property, data are classified to produce maps that retain as much of the spatial structure of the raw data as possible. Such a classification retains as much visual complexity as possible for a given number of class intervals unlike the traditional quantile classification, the only other classification based solely on ordinal relationships.